Advanced search
2 files | 452.27 KB Add to list

Robust language recognition via adaptive language factor extraction

Brecht Desplanques (UGent) , Kris Demuynck (UGent) and Jean-Pierre Martens (UGent)
Author
Organization
Abstract
This paper presents a technique to adapt an acoustically based language classifier to the background conditions and speaker accents. This adaptation improves language classification on a broad spectrum of TV broadcasts. The core of the system consists of an iVector-based setup in which language and channel variabilities are modeled separately. The subsequent language classifier (the backend) operates on the language factors, i.e. those features in the extracted iVectors that explain the observed language variability. The proposed technique adapts the language variability model to the background conditions and to the speaker accents present in the audio. The effect of the adaptation is evaluated on a 28 hours corpus composed of documentaries and monolingual as well as multilingual broadcast news shows. Consistent improvements in the automatic identification of Flemish (Belgian Dutch), English and French are demonstrated for all broadcast types.
Keywords
language factor extraction, model adaptation, language recognition, iVectors

Downloads

  • LID2 camera ready.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 205.88 KB
  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 246.40 KB

Citation

Please use this url to cite or link to this publication:

MLA
Desplanques, Brecht, Kris Demuynck, and Jean-Pierre Martens. “Robust Language Recognition via Adaptive Language Factor Extraction.” 15th Annual Conference of the International Speech Communication Association, Proceedings. 2014. 2160–2164. Print.
APA
Desplanques, B., Demuynck, K., & Martens, J.-P. (2014). Robust language recognition via adaptive language factor extraction. 15th Annual Conference of the International Speech Communication Association, Proceedings (pp. 2160–2164). Presented at the 15th Annual Conference of the International Speech Communication Association (Interspeech 2014).
Chicago author-date
Desplanques, Brecht, Kris Demuynck, and Jean-Pierre Martens. 2014. “Robust Language Recognition via Adaptive Language Factor Extraction.” In 15th Annual Conference of the International Speech Communication Association, Proceedings, 2160–2164.
Chicago author-date (all authors)
Desplanques, Brecht, Kris Demuynck, and Jean-Pierre Martens. 2014. “Robust Language Recognition via Adaptive Language Factor Extraction.” In 15th Annual Conference of the International Speech Communication Association, Proceedings, 2160–2164.
Vancouver
1.
Desplanques B, Demuynck K, Martens J-P. Robust language recognition via adaptive language factor extraction. 15th Annual Conference of the International Speech Communication Association, Proceedings. 2014. p. 2160–4.
IEEE
[1]
B. Desplanques, K. Demuynck, and J.-P. Martens, “Robust language recognition via adaptive language factor extraction,” in 15th Annual Conference of the International Speech Communication Association, Proceedings, Singapore, Singapore, 2014, pp. 2160–2164.
@inproceedings{5713973,
  abstract     = {This paper presents a technique to adapt an acoustically based
language classifier to the background conditions and speaker
accents. This adaptation improves language classification on
a broad spectrum of TV broadcasts. The core of the system
consists of an iVector-based setup in which language and channel
variabilities are modeled separately. The subsequent language
classifier (the backend) operates on the language factors,
i.e. those features in the extracted iVectors that explain the observed
language variability. The proposed technique adapts the
language variability model to the background conditions and
to the speaker accents present in the audio. The effect of the
adaptation is evaluated on a 28 hours corpus composed of documentaries and monolingual as well as multilingual broadcast
news shows. Consistent improvements in the automatic identification
of Flemish (Belgian Dutch), English and French are demonstrated for all broadcast types.},
  author       = {Desplanques, Brecht and Demuynck, Kris and Martens, Jean-Pierre},
  booktitle    = {15th Annual Conference of the International Speech Communication Association, Proceedings},
  keywords     = {language factor extraction,model adaptation,language recognition,iVectors},
  language     = {eng},
  location     = {Singapore, Singapore},
  pages        = {2160--2164},
  title        = {Robust language recognition via adaptive language factor extraction},
  url          = {http://www.isca-speech.org/archive/interspeech_2014/i14_2160.html},
  year         = {2014},
}